INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

RBP9 is a Drosophila RNA binding protein that shares a high level of sequence similarity with Drosophila ELAV (41) and human Hu proteins (HuC, HuD, Hel-N1, and HuR) (26, 29, 48). Proteins in this family are known to be expressed in the nuclei of neuronal cells right after the completion of mitotic division. RBP9 is expressed predominantly in the nuclei of cells of the central nervous system (CNS), after the CNS metamorphosis that occurs during the pupal period (19). The related human Hu proteins are also expressed primarily in neurons and are localized preferentially in the nuclei (8). Hu proteins are absent in neuroblasts but appear in subsequent early-lineage neurons and maturing neuronal cells. Thus, it has been suggested that proteins in this family are required for neuronal maturation.

A role for Rbp9 in neurogenesis is further suggested by the fact that ELAV is expressed specifically in the nuclei of all neurons (41), and loss-of-function alleles of elav are embryonic lethal, causing abnormal CNS development (40). Recently, elav was suggested to regulate neuron-specific splicing of neuroglian pre-mRNA (24), which is consistent with the presence of RNA binding motifs in the ELAV protein and its nuclear localization pattern.

However, in vitro studies suggest mRNA stability rather than pre-mRNA splicing as a functional target of the Hu proteins. For example, Hu proteins were shown to bind to stretches of U residues (the AU-rich element), and this interaction increases the stability of the bound reporter mRNAs in a cell culture system (5, 9, 21, 26, 28, 32, 37). Given that Hu protein is localized mainly in the nucleus, the cytoplasmic function of mRNA stabilization appears to be accomplished by the shuttling of nuclear Hu proteins to the cytoplasm (9). Because Hu protein binding sequences are often found in mRNAs that encode cell growth regulators, it has been suggested that the Hu proteins control cell proliferation by regulating the stability of mRNAs that encode cell proliferation and/or differentiation signal proteins (3, 4, 5, 15, 21, 26, 32). For example, the in vitro binding of Hel-N1 protein to the 3' untranslated region (UTR) of Id mRNA, which encodes a transcriptional repressor abundant in undifferentiated neuronal precursor cells, suggests the involvement of Hel-N1 in the regulation of nerve cell development (21). Recently, we found that RBP9 down regulates the expression of extramacrochaetae (emc), a Drosophila homolog of Id, possibly through the interaction of RBP9 protein with emc mRNA (36). These results indicate that the RBP9 family of proteins binds to mRNAs and functions in the regulation of cell growth and differentiation.

In addition to the two putative functions of the Hu protein gene family discussed above, the expression of certain Hu proteins in nonneuronal tissues indicates the existence of additional physiological functions. In vertebrates, four closely related Hu homologs are expressed in a distinct developmental pattern. For example, in adult frogs, elrC and elrD are expressed exclusively in nerve cells during specific developmental stages, whereas elrA is expressed in all tissues throughout development. In particular, elrB is expressed in testis and ovaries, in addition to its stage-specific expression in the brain (13). Therefore, each of the related Hu homologs appears to participate in the regulation of distinct developmental processes.

In this paper, we present the results of experiments designed to decipher the function(s) of RBP9. We show that RBP9 protein is expressed in the cytoplasm of ovaries, as well as in the nuclei of neuronal cells. Analysis of Rbp9 mutants revealed that although mutant cystocytes divide continuously, their differentiation is arrested. In addition, we show that RBP9 protein binds specifically to the U-rich region of bag-of-marbles (bam) RNA in vitro and that BAM protein expression is expanded in Rbp9 mutant ovaries. The specific interaction between RBP9 protein and bam mRNA in vitro suggests that RBP9 regulates either the stability or the translational competence of bam mRNA, which in turn affects cystocyte differentiation of ovaries.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Determination of the P-element insertion site at the 23C region. The DNA sequences flanking the P elements at the 23C region of mutant Drosophila P944, P1046, l(2)K00632, l(2)K05431, l(2)K05432, l(2)K07920, l(2)K12901, l(2)K16124, and l(2)K08224 strains (49) were amplified by inverse PCR (7). Each amplified fragment from P-element insertion lines was cloned, labeled with ³²P, and hybridized to dot-blotted DNA from the following cosmid contigs: 166H10, 16A8, 82F12, 78F3, 193E7, 122A10, 25A6, and 132H9 (44). To determine the exact location of a P element within the cosmid clones that gave a positive signal in the dot blot hybridization analysis, the cosmid DNA was digested, blotted on a nitrocellulose membrane, and probed with a P-element flanking sequence. The exact location of the P-element insertion site was confirmed by directly sequencing the amplified P-element flanking sequences as described by Rao (38).

P-element-mediated mutagenesis. The l(2)K12901 strain has one P element that contains a mini-white gene (P[w⁺]) at the 23C region. To mobilize the P element of l(2)K12901, we crossed l(2)K12901/CyO flies with w/w; TMS, Sb P[ry⁺ 2-3]/Dr flies (39), which harbor a helper P element (2-3) on the third chromosome, and isolated male progeny that had both P elements. After the helper P element provided the necessary transposase to mobilize P[w⁺] to new insertion sites, we crossed four of these males with four w¹¹¹⁸ homozygous females in each vial (total number of vials = 2,500) to eliminate the P[ry⁺ 2-3] element. Progeny with eye colors different from those of the parents were screened for mobilized P elements. No more than one fly with changed eye color was selected from each vial and crossed with w¹¹¹⁸ virgin females to establish lines that carried P elements that had moved to new insertion sites. We established 900 lines from this screening, and the location of each P-element insertion site was determined by PCR. A PIR primer (7) complementary to the inverted repeats of the P element was used with eight 20-nucleotide Rbp9-specific primers for the PCR. These eight Rbp9-specific primers are referred to as 199, 2109, 3219, 5125, 6710, 8370, 9909, and 11325. The numbers indicate the nucleotide position of the 5' end of each primer in the Rbp9 genomic sequence (GenBank accession no. S55886), and these eight primers corresponded to nearly every 1- to 2-kb region throughout the entire gene.

Rbp9¹

Rbp9¹

Rbp9¹

Immunoblot analysis. Five adult flies or the dissected ovaries and the carcass from an equivalent number of flies were homogenized with a disposable microdouncer in 30 µl of protein sample loading buffer in a microcentrifuge tube. The fly debris was removed by centrifugation (15,000 × g, 5 min). Protein extracts (~4 µg of protein) were subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and transferred to a nitrocellulose membrane for Western analysis (42).

Immunostaining of ovaries. Ovaries (10 to 20 pairs) from healthy young flies were dissected in 1× phosphate-buffered saline (PBS) and fixed for 20 min in 4% paraformaldehyde in 1× PBS. The fixed ovaries were washed several times over a 2-h period with PBT (0.1% Tween 20 in 1× PBS). Anti-RBP9 antibody (Ab) was raised against the C-terminal half of the protein and affinity-purified as described elsewhere (19). Abs were diluted (RBP9 Ab, 1:50; cytoplasmic BAM Ab [30], 1:200; and HTS [hu-li tai shao protein] Ab [51], 1:5) in TNBTT (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.1% bovine serum albumin, 0.1% Triton X-100, 0.05% Thimerosal [Sigma, St. Louis, Mo.]), and the ovaries were incubated in an Ab-containing solution at 4°C overnight. After being washed once with TNBTT, ovaries were incubated with TNBTT containing 2% goat serum for 30 min, washed with TNBTT for 2 h, and then incubated with fluorescein isothiocyanate-conjugated secondary immunoglobulin G (Sigma) in TNBTT at a 1:100 dilution for 2 h. After washing with PBS, ovaries were mounted with Vectashield (Vector Laboratories, Burlingame, Calif.) mounting medium. The RBP9 Ab was used after incubation with ELAV-agarose (0.1 ml) for 4 h at 4°C to remove any ELAV-cross-reacting material.

UV cross-linking assay. Protein-DNA UV cross-linking assays were performed as described elsewhere (20). Recombinant RBP9 (rRBP9; 60 ng) (36) was preincubated for 10 min with 10 µg of yeast tRNA in a 10-µl reaction mixture that contained 1 µl of 10× reaction buffer A (32 mM MgCl₂, 20 mM ATP, 1 mg of bovine serum albumin per ml, 60 mM HEPES-KOH [pH 7.9]). The 3' UTR of bam mRNA (450 bp between the termination codon and the polyadenylation signal) was prepared by PCR amplification of wild-type Drosophila cDNA with primers bam5 (5'-TTT CTA GAA CTA ATG CTG TGC ACA TCG AT-3') and bam3 (5'-TTT CTA GAT GAC TTT CAA AAT ACA AAT G-3'). The amplified fragment, which was digested with XbaI, was cloned into the XbaI site of pBluescript KS+II (Stratagene) to make pKSBam3UTR, and the insert was sequenced to confirm the absence of a mutation. An RNA probe encoding the bam 3' UTR (containing three putative RBP9 binding sites [RBP9BS] at 274, 371, and 388 bp from the termination codon) was transcribed by T7 RNA polymerase in the presence of [³²P]UTP from the pKSBam3UTR template that has been linearized with BamHI. The ³²P-labeled RNA probe (100 fmol) was added to the reaction mixture, and the sample was incubated for an additional 10 min at room temperature. The sample was placed on ice and irradiated with UV light (10⁵ ergs/mm²) with use of a Stratagene (La Jolla, Calif.) UV cross-linker. The RNA was digested with RNase A (30 µg) for 15 min at 37°C and mixed with protein gel loading buffer. Samples were boiled for 90 s and subjected to SDS-PAGE and autoradiography. For UV cross-linking competition assays, a 20- to 400-fold excess of competitor RNA oligonucleotides was added to the reaction mixtures together with the ³²P-labeled RNA probes. The competitor RNA oligonucleotide RBP9BS (sense strand) was designed on the basis of the consensus binding sequences of rRBP9 (UUUXUUUU) (36) and Hel-N1 (RWUUUAUUUWR) (10), which were identified with the use of SELEX (25). The RNA oligonucleotides used for these assay were RBP9BS sense (5' UUG AUU UAU UUU GAU UUU AUU UAG UU 3') and RBP9BS antisense (5' GAA AAA AAA AGA AAA AAA AAA GAA 3').

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of RBP9 in the cytoplasm of early developing cystocytes during oogenesis. To examine whether the Drosophila Hu homolog RBP9 is expressed in germ line cells, we dissected ovaries from wild-type flies and examined the expression of RBP9 in both the ovaries and the carcass. Western analysis showed that RBP9 was expressed in ovaries as well as in the carcass where the CNS is retained (Fig. 1A).

View larger version (64K): [in this window] [in a new window]   FIG. 1.   Expression and localization of RBP9 in the ovaries. (A) Expression of RBP9 protein in ovaries. Ovaries were dissected from wild-type flies, and whole-cell extracts were prepared from the dissected ovaries (O) and carcass (C). The whole-cell extracts were resolved in an SDS-polyacrylamide (10%) gel and immunoblotted with affinity-purified anti-RBP9 antibody. As a protein loading control, the same blot was probed with an antibody to yeast TATA binding protein (TBP); the cross-reacting protein bands are shown. (B) Cytoplasmic localization of RBP9 protein in wild-type ovaries. Ovaries were stained with anti-RBP9 antibodies and analyzed by confocal microscopy. The ovariole in the middle shows, from left to right, the germarium regions (1, 2a, 2b, and 3) and stage 2 egg chamber (S2). The oocytes at the posterior of each egg chamber are marked with arrows. In the stage 3 egg chamber (S3), RBP9 staining is detected only in the oocyte.

Isolation of Rbp9 mutant alleles by P-element-induced mutagenesis. The distinct localization pattern of RBP9 in the cytoplasm of ovaries, in addition to the nuclei of cells of the CNS, provided a unique opportunity to examine both the cytoplasmic and nuclear functions of Rbp9 during development. These observations prompted us to isolate a collection of Rbp9 mutants. Because our initial attempt to isolate Rbp9 mutants by using ethyl methanesulfonate treatment was not successful (18), we used a local P-element mutagenesis protocol (see Materials and Methods). As a first step, we screened existing fly strains that had P elements inserted at the 23C region for one that contained a P element close to the Rbp9 locus. Southern blot analyses performed on Rbp9 genomic clones with the P-element flanking sequences as probes revealed that the l(2)K12901 strain contains a P element inserted within the Rbp9 gene. Sequence analysis of the P-element flanking regions confirmed its location at the Rbp9 intron upstream of the third promoter (Fig. 2).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Locations of P-element insertions in Rbp9 mutant alleles isolated by P-element-induced mutagenesis. P1, P2, and P3 indicate the three alternative Rbp9 promoters; the splicing patterns of alternatively used exons are shown with lines. The P-element insertion sites of I(2)K12901 and Rbp9P[1374] flies used to initiate P-element local mutagenesis, the ultimate locations of the P elements in the 87 lines isolated, the four Rbp9 mutant alleles (P[2567], P[2398], P[2690], and P[2775]) described in this paper are indicated, and the locations of primers used to map the insertion sites of P elements are indicated. The number above each arrow at the top indicates the 5' end of the primer corresponding to the nucleotide number of Rbp9 as described by Kim and Baker (19).

View larger version (15K): [in this window] [in a new window]   FIG. 3.   Immunoblot analysis of Rbp9 alleles. (A) Crude fly extracts (4 µg) were analyzed by Western blotting with anti-RBP9 Ab. The fly strains analyzed are indicated above the lanes (w1118 [w] was used as wild type). (B) The expression of RBP9 protein in ovaries (O) and carcass (C) is shown for wild-type and Rbp9P[2567] female flies. As a loading control, the same blot was probed with an Ab to yeast TATA binding protein (TBP); the cross-reacting protein band in each lane is shown.

Female-specific sterility of Rbp9 mutants. Although the predominant expression of RBP9 in the adult CNS suggested a putative function of RBP9 in CNS development, we could detect no obvious defect in mutant viability, CNS development, or behavior. However, each Rbp9 mutant showed some degree of defects in fertility. Rbp9^P[2690] homozygotes were completely sterile, while Rbp9^P[2775] and Rbp9^P[2398] homozygotes showed a reduction in fertility. The numbers of progeny eclosed from Rbp9^P[2690], Rbp9^P[2775], and Rbp9^P[2398] homozygous parents were 0, 3, and 23%, respectively, of those eclosed from wild-type parents (Table 1). When homozygous females of each of the Rbp9 mutant alleles were crossed to wild-type males, the degree of sterility observed was similar to that of each homozygote. On the contrary, none of the mutant males showed a defect in fertility when crossed to wild-type females (Table 1). Therefore, Rbp9 is required for fertility only in females.

Rbp9¹

Rbp9¹

Defective ovary development in Rbp9 mutants. To understand how the loss of Rbp9 activity caused female sterility, we examined the phenotypes of ovaries isolated from Rbp9 mutant strains. In whole ovaries examined by nuclear (4',6-diamidino-2-phenylindole [DAPI]) staining (Fig. 4), ovarioles from strains bearing the Rbp9 alleles that showed the most severe sterility phenotype (Rbp9^P[2690] and Rbp9¹) were filled with egg chambers containing abnormal number of cystocytes that never developed beyond a stage 6 egg chamber (Fig. 4A and B and data not shown). In addition, the germarium regions were enlarged as often detected in ovarian tumors (11, 23, 31, 35, 43). In a few cases the ovarioles were completely devoid of egg chambers. As the Rbp9 mutant flies aged, we observed that the underdeveloped egg chambers became filled with numerous small cells (the "tumorous bag" phenotype).

View larger version (135K): [in this window] [in a new window]   FIG. 4.   Allele-specific phenotypes of Rbp9 mutant ovaries. The anterior end of the ovaries is oriented at the right side of each panel except panel F, where the anterior end is oriented toward the top. Ovaries consist of ovarioles in which progressively more advanced stages of oogenesis are arranged linearly from anterior (right or top in panel F) to posterior (left or bottom in panel F). (A to F) DAPI staining patterns of ovaries from wild-type and mutant flies. (A) Wild-type ovaries. Each egg chamber consists of one oocyte and 15 nurse cells surrounded by follicle cells. (B to E) Ovaries from Rbp9P[2690], Rbp9P[2775], Rbp9P[2398], and Rbp9P[2567] homozygous flies, respectively. (F) Ovaries from an Rbp9P[2398]/Df hemizygous fly. Egg chambers with fewer or more than 15 nurse cells are indicated with short or long arrows, respectively. Black arrowheads represent egg chambers with tumorous cells.

Requirement of Rbp9 for oocyte determination and positioning. In addition to the function of Rbp9 in the regulation of germ cell proliferation and differentiation, the specific localization of RBP9 protein in the oocyte of the early egg chambers suggests that RBP9 is also required for proper oocyte determination. Because the strong Rbp9 mutant alleles (Rbp9^P[2690] and Rbp9¹ alleles) did not produce egg chambers developed sufficiently to allow study of the later stage-specific function of Rbp9, we examined the oocyte determination process of hypomorphic Rbp9 (Rbp9^P[2398] and Rbp9^P[2567]) mutants, which contain normally shaped egg chambers. In wild-type ovaries, oskar (osk) mRNA is concentrated in the oocytes located at the posterior end of each egg chambers (Fig. 5A). However, in the Rbp9^P[2398] mutant ovaries, osk mRNA often accumulated in more than one cell (Fig. 5B). Even when osk mRNA did accumulate in only one cell, it often was not positioned at the most posterior end of the egg chambers. This alteration in the distribution pattern of osk mRNA demonstrated that Rbp9 function is also required for proper oocyte determination and positioning.

View larger version (122K): [in this window] [in a new window]   FIG. 5.   Pattern of osk mRNA distribution in wild-type (A) and Rbp9P[2398] (B) ovaries. The hybridization signal appears as darkly staining material. Arrowheads indicate egg chambers that either contain an osk mRNA-accumulating cell that is not located at the posterior of egg chamber or have more than one cell that accumulates osk mRNA. Progressively older egg chambers are shown from left to right.

Requirement of Rbp9 function for regulation of cystocyte differentiation. Immunostaining of wild-type ovaries with antibodies to HTS (51), a fusome component, initially shows two stem cells with dot-shaped spectrosomes; as the cystocytes divide, fusomes with branched structures are observed. The branched fusome structure is observed most clearly when cystocytes lose the rosette conformation after the fourth round of mitotic division and take on a lens-shaped appearance in the most anterior section of germarium region 2 (Fig. 6A). This cyst development requires a number of genes, including ovarian tumors (otu), benign gonial cell neoplasm (bgcn), and bam (11, 12, 46). Analysis of the germarium regions of bgcn or bam mutants with antibodies to fusome components reveals that most of the cystocytes have dot-shaped or single-branched fusomes, whereas the cystocytes of otu mutant ovaries have branched fusomes despite the similar tumorous phenotype (27, 31, 47).

View larger version (77K): [in this window] [in a new window]   FIG. 6.   Immunostaining of the germarium with Abs to HTS (A and B) or BAM (C to E). Wild-type (A and C) and Rbp91 mutant (B, D, and E) ovaries were stained with Abs to the cytoplasmic BAM and HTS proteins. The anterior ends of the germarium are at the left of each panel. (A) At the tip of the germarium, a fusome in the stem cell appears as a spherical dot (arrowhead). As cystocytes divided, the fusome becomes elongated and finally appears as a branched structure (long arrow). (B) Most of fusomes in the Rbp9 mutant appear to be branched (short arrow) but do not have the shape of a fully branched structure. Some cystocytes contain fusomes in the shape of spherical dots without an obvious visible connection between each other (*). (C) Cytoplasmic staining of BAM protein is clearly visible in a cluster of four cystocytes located near the anterior end of the wild-type germarium. (D) Rbp9 mutant germarium contains an increased number of cells that express BAM proteins. Several clusters of cystocytes are stained with BAM Ab. (E) BAM protein is still detectable even when the cystocytes show a lens-shaped morphology, which is usually detected in germarium region 2b (arrow).

Rbp9¹

Rbp9¹

Binding of RBP9 to the 3' UTR of bam mRNA in vitro. Our observation that an increased number of Rbp9 mutant cystocytes express BAM protein suggests that Rbp9 may be required for the down regulation of BAM expression. For example, RBP9 may bind to bam mRNA and regulate its stability or translatability. In vitro binding assays identified the sequence UUUAUUU as an RBP9 consensus binding site. Examination of the nucleotide sequence of bam mRNA by using the WINDOW program (Wisconsin Package; GCG Inc.) identified three UUUAUUU sequences within the 3' UTR, located 177, 80, and 63 bp upstream of the poly(A) signal. To test whether these repeats represent authentic RBP9BS, we used a UV cross-linking assay to detect RBP9 binding to the bam mRNA 3' UTR. As shown in Fig. 7B, the 3' UTR of bam mRNA cross-linked efficiently to RBP9 protein (lane 2). To confirm the specificity of binding, we tested the ability of sense and antisense RNAs to compete for RBP9 binding. When added to the binding reaction, sense competitor RNA oligonucleotide that contained two repeats of the RBP9BS efficiently competed with the bam 3' UTR probe for RBP9 binding (lanes 3 to 5). Conversely, antisense RNA that contained the complementary RBP9BS sequences did not compete efficiently for RBP9 binding, even when a 400-fold excess was added to the binding reaction (lanes 6 to 8). These results indicate that RBP9 binds specifically to the bam mRNA 3' UTR and suggest that RBP9 may down regulate BAM expression by interacting with the AU-rich elements of bam mRNA.

View larger version (27K): [in this window] [in a new window]   FIG. 7.   Binding of RBP9 protein to the U-rich elements within the 3' UTR of bam mRNA. (A) locations of the UUUAUUU motifs () in the 3' UTR of bam mRNA. The second and third Rbp9 binding motifs () are located 9 bp apart near the poly(A) signal. The probe used for the UV cross-linking assay is shown. The locations of the bam coding region and poly(A) tail (AAAAA) are marked. (B) UV cross-linking of RBP9 protein to bam mRNA. Recombinant RBP9 protein (60 ng) was UV cross-linked to the 3' UTR of bam mRNA. RBP9BS sense (lanes 3 to 5) and antisense (lanes 6 to 8) RNA oligonucleotides were added as competitor RNAs to the binding reaction in 40-, 120-, and 400-fold excesses, respectively.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Because the RNA binding proteins RBP9, ELAV, and Hu were identified originally in nerve cell nuclei, researchers suggested that they function in neurospecific pre-mRNA processing (19, 29, 48). However, the results of subsequent studies on Hu proteins suggested that they play a role in the regulation of mRNA stability in the cytoplasm of developing neuronal cells (5, 9, 21, 26, 28, 32, 37). Although Hu proteins were shown recently in a tissue culture system to increase the stability of reporter mRNAs that contain Hu protein binding sequences (9), direct evidence of such a cytoplasmic function obtained under physiological conditions was lacking. In this study, we show that RBP9 is localized in the cytoplasm and regulates BAM expression in differentiating germ cells. These findings strongly support the idea that the RBP9 family proteins regulate mRNA stability and/or translational competence of mRNAs that encode cellular differentiation signal proteins.

In addition to the proposed cytoplasmic function of Rbp9, the detection of Rbp9 in a tissue outside the CNS is intriguing. Although elrA, elrB, and HuR were found in cells other than neurons, their physiological roles in these nonneuronal cells have not been established (13, 29). In this study, genetic, molecular, and morphological examination of Rbp9 mutants revealed a requirement for Rbp9 in the regulation of female germ cell differentiation. The absence of RBP9 protein in the germarium caused cystocytes to be arrested at various undifferentiated stages, and tumorous ovarioles filled with the accumulating cystocytes were formed. This defect seems to be mediated through bam, as RBP9 interacts with bam mRNA in vitro and BAM expression is expanded in the Rbp9 mutants. Moreover, the unusual expression of BAM in germarium region 2b cystocytes suggests that the expansion of BAM expression is not just a simple reflection of the increase of early stage cystocytes in the tumorous ovarioles. Rather, we suggest that down regulation of BAM expression is required for proper cystocyte differentiation and that this process is not properly regulated in the Rbp9 mutants.

It has been suggested that bam serves multiple functions in a number of the developmental stages of oogenesis. A loss of function mutation in bam causes female cysts to proliferate like stem cells (30, 31), whereas ectopic expression of bam in ovaries, including the stem cells, eliminates oogenic germ line stem cells (33). Therefore, bam appears to function in suppressing stem cell fate in stem cell daughters and to promote cystoblast differentiation (33). Analysis of encore mutants suggested additional functions for bam during the later stages of cystocyte development. BAM is expressed only in the dividing cystoblast and cystocytes and is removed from the cystocytes after four rounds of mitotic division (30). As the expansion of bam transcripts is correlated with one extra round of mitosis in the encore mutant germarium, bam may act as part of a titratable counting mechanism whose levels determine the number of mitotic divisions (14). Although both Rbp9 and encore are involved in the regulation of cystocyte proliferation and differentiation through their interactions with bam, they have quite different mutant phenotypes (reference 14 and this study) and, thus, appear to act by distinct mechanisms. Because RBP9 binds to the 3' UTR of bam transcripts in vitro, RBP9 may be involved in the destabilization of residual bam transcripts or in the process of making bam transcripts unsuitable for translation. This is consistent with our recent finding that the level of emc mRNA, which also interacts with RBP9 in vitro, is down regulated by Rbp9 in vivo (36).

The observations that RBP9 protein is concentrated in the oocyte of wild-type early-stage egg chambers and that mispositioned or abnormal numbers of oocytes were detected in hypomorphic Rbp9 mutants suggest that Rbp9 is required not only for cystocyte differentiation but also for oocyte development. This process could be mediated through a target RNA(s) other than the bam transcript. Because RBP9 binds to a rather simple RBP9 consensus binding sequence in vitro and this sequence is found in many transcripts, a number of RNAs expressed in the ovary have the potential to interact with RBP9 (36). However, the observation that all transcripts that contain the RBP9 consensus binding sequences are not regulated by RBP9 (36) suggests that additional cis- or trans-acting components must be required for the functional interaction of RBP9 with its specific target RNAs in vivo. Therefore, the identification of these additional factors is necessary to elucidate the mechanism by which RBP9 regulate its target RNAs.

Rbp9 appears to function in at least one more process during oogenesis. Although we did not observe gross egg chamber defects by DAPI staining in weak Rbp9 alleles (Rbp9^P[2398] and Rbp9^P[2567]), the fertility of these alleles measured by the number of eclosed adult flies remained low. This reduction in fertility can be explained in two ways. First, the overall number of eggs laid by the mutants was smaller than that laid by wild-type flies (data not shown). Second, eggs laid by the weak Rbp9 mutant flies were somewhat defective in axis formation. About 20% of the mutant embryos showed body patterning defects when determined by cuticle preparation. Posterior body patterning defects (45) were most often detected, while some embryos showed a dorsalized phenotype (data not shown). Therefore, Rbp9 appears to be required for multiple distinct developmental steps during oogenesis.

Although we detected several oogenesis defects in Rbp9 mutant flies, we observed no defect in CNS function. Serial sections of adult flies carrying either an Rbp9 null allele or wild-type Rbp9 were stained with DAPI, anti-RBP9 Ab, and monoclonal Ab (MAb) 22C10 (52) in order to visualize nuclei, RBP9, and neural cytoplasmic antigen, respectively. The nuclear staining patterns showed no morphological differences between wild-type and mutant flies, despite the lack of RBP9 protein in the mutant CNS (data not shown). Neurons from Rbp9 mutant strains expressed neuron-specific antigen, and axons stained normally with MAb 22C10. The absence of an effect on the nervous systems of Rbp9 mutant flies was extended further in that no obvious abnormalities were detected in flight, feeding, or mating behaviors (data not shown). The coexpression of ELAV and RBP9 in the adult CNS may provide functional redundancy.

Because perturbation of germ line sex determination has been reported with mutations that give rise to the tumorous ovarian phenotype (34), we examined the sex-specific pattern of Sex lethal (Sxl) pre-mRNA splicing in Rbp9 mutants with respect to the presence of the male-specific Sxl exon. The male-specific Sxl transcript was not observed in wild-type or Rbp9^P[2567] females (data not shown). However, Rbp9^P[2775] females had a small amount and Rbp9^P[2690] homozygous females had a higher concentration of the male-specific Sxl transcript (data not shown). The observation that greater amounts of the male-specific Sxl transcript are produced in mutants with a more severe ovarian tumor phenotype implies that germ line sex determination is correlated with the cystocyte differentiation process of the germarium. Because the initial steps of the cystocyte differentiation process are very similar for oogenesis and spermatogenesis, the arrest of cystocyte development prior to the differentiation of egg chambers may send an erroneous signal to the cystocyte in the germarium causing the cells to follow a male-like developmental pathway.

Although Rbp9 is expressed mainly in postmitotic cells both in the CNS and in the ovaries, the distinct cellular localization patterns of RBP9 protein in each tissue type suggests a unique function for RBP9 in each location. SXL is another example of an RNA binding protein with distinct tissue-specific functions. SXL was originally identified as an alternative-splicing regulator in the somatic sex determination pathway (1, 6). Despite the presence of a considerable amount of cytoplasmic SXL protein in certain tissues, only the nuclear protein was considered to be functional. However, it was later shown that cytoplasmic SXL protein binds to the UTR of msl-2 mRNA and inhibits its translation (2, 17). Although SXL executes two different functions in two distinct cellular locations, both processes are mediated by the recognition of similar polypyrimidine elements that are found in the UTRs of msl-2 and the introns of genes involved in sex determination. The localization of RBP9 and SXL proteins in two distinct cellular compartments could be guided by their association with tissue-specific localization factors or by modification of their phosphorylation status. An understanding of the cellular processes that regulate protein localization should aid us in deciphering the mechanisms used by RBP9 and SXL in their various modes of developmental regulation.